结肠压力数据的自组织聚类分析

Clustering Analysis of Colonic Pressure Data by Self Organization Feature Maps

结肠测压对结肠运动障碍疾病 ,如慢传输型便秘 (Slow Transit Constipation,STC)的诊断有较高的价值 ,但由于它是介入性检查 ,实施起来很不方便 .为此 ,对实际采集的结肠压力信号进行特征提取 ,采用自组织特征映射的方法进行结肠压力信号的自聚类和识别 .结果表明 ,这种方法可以区分正常与异常的结肠压力信号 ,与初始的医疗诊断相符合 ,可以作为判断肠道运动性能的辅助诊断手段 ,为根据结肠压力信号进一步识别肠道的动力性能、诊断 STC奠定了基础 .

Colonic manometry is a widely used method which estimates functions of the alimentary tract and is most valuable to the diagnosis of colonic functions. But the measure is not convenient because of the intervening of the manometry and some unavoidable objective factors. This paper dealt with the features extracted from the colonic pressure data and makes a clustering analysis of the colonic pressure data by using self organization feature maps (SOFM). The data are classified as normal and abnormal which accords with the original diagnosis. This method can serve as an assistant diagnosis of the colonic motility and contributes to further research of colonic motility based on pressure data.